# Required Packages
import pandas as pd
import numpy as np
import pickle

# preprocessing
from sklearn.impute import SimpleImputer

# Visualisation libraries

## Text
from colorama import Fore, Back, Style
from IPython.display import Image, display, Markdown, Latex, clear_output

## progress bar
import progressbar

## plotly
from plotly.offline import init_notebook_mode, iplot 
import plotly.graph_objs as go
import plotly.offline as py
from plotly.subplots import make_subplots
import plotly.express as px

## seaborn
import seaborn as sns
sns.set_style("whitegrid")
sns.set_context("paper", rc={"font.size":12,"axes.titlesize":14,"axes.labelsize":12})

## matplotlib
import matplotlib.pyplot as plt
from matplotlib.font_manager import FontProperties
from matplotlib.patches import Ellipse, Polygon
import matplotlib.gridspec as gridspec
import matplotlib.colors
from pylab import rcParams
from matplotlib.font_manager import FontProperties
from mpl_toolkits.axes_grid1.inset_locator import inset_axes
plt.style.use('seaborn-whitegrid')
import matplotlib as mpl
mpl.rcParams['figure.figsize'] = (17, 6)
mpl.rcParams['axes.labelsize'] = 14
mpl.rcParams['xtick.labelsize'] = 12
mpl.rcParams['ytick.labelsize'] = 12
mpl.rcParams['text.color'] = 'k'
%matplotlib inline

import warnings
warnings.filterwarnings("ignore")

Home / Business Analytics and Modeling

Starbucks Offer Personalizations

• Preprocessing
• Exploratory Data Analysis
• Modeling: Multi Output Random Forest Classifier
• Modeling: Keras Multi-layer Perceptron (MLP) for Multi-Class Classifications

Starbucks Offer Personalizations

In this article, we investigate a set simulated dataset that mimics customer behavior on the Starbucks rewards mobile app. Starbucks tends to send out offers to users of the mobile app once every few days. These offers are exclusive, that is not all users receive the same offer. An offer can contain a discount for their products or sometimes BOGO (buy one get one free). These offers have a validity period before the offer expires. The article here is inspired by a towardsdatascience.com article.

Loading the Datasets

def Header(Text, L = 100, C = 'Blue', T = 'White'):
    BACK = {'Black': Back.BLACK, 'Red':Back.RED, 'Green':Back.GREEN, 'Yellow': Back.YELLOW, 'Blue': Back.BLUE,
         'Magenta':Back.MAGENTA, 'Cyan': Back.CYAN}
    FORE = {'Black': Fore.BLACK, 'Red':Fore.RED, 'Green':Fore.GREEN, 'Yellow':Fore.YELLOW, 'Blue':Fore.BLUE,
         'Magenta':Fore.MAGENTA, 'Cyan':Fore.CYAN, 'White': Fore.WHITE}
    print(BACK[C] + FORE[T] + Style.NORMAL + Text + Style.RESET_ALL + ' ' + FORE[C] +
          Style.NORMAL +  (L- len(Text) - 1)*'=' + Style.RESET_ALL)
    
def Line(L=100, C = 'Blue'):
    FORE = {'Black': Fore.BLACK, 'Red':Fore.RED, 'Green':Fore.GREEN, 'Yellow':Fore.YELLOW, 'Blue':Fore.BLUE,
         'Magenta':Fore.MAGENTA, 'Cyan':Fore.CYAN, 'White': Fore.WHITE}
    print(FORE[C] + Style.NORMAL + L*'=' + Style.RESET_ALL)
    
# Portfolio Dataset
Header('Portfolio Dataset:')
Portfolio = pd.read_csv('StarBucks/Portfolio_Clean.csv')
display(Portfolio.head().style.hide_index())

# Profile Dataset
Header('Profile Dataset:', C = 'Green')
Profile = pd.read_csv('StarBucks/Profile_Clean.csv')
display(Profile.head().style.hide_index())

# Transcript Dataset
Header('Transcript Dataset:', C = 'Red')
Transcript = pd.read_csv('StarBucks/Transcript_Clean.csv')
display(Transcript.head().style.hide_index())
Line()

User_Data = pd.read_csv('StarBucks/User_Data.csv')
Data = pd.read_csv('StarBucks/Data.csv')

# Dictionaries
with open('StarBucks/Feat_Dict.pkl', 'rb') as fp:
    Feat_Dict = pickle.load(fp)

Portfolio Dataset: =================================================================================

Reward	Difficulty	Duration	Offer Type	Offer ID	Email	Mobile	Social	Web
10	10	7	BOGO	ae264e3637204a6fb9bb56bc8210ddfd	1	1	1	0
10	10	5	BOGO	4d5c57ea9a6940dd891ad53e9dbe8da0	1	1	1	1
0	0	4	Informational	3f207df678b143eea3cee63160fa8bed	1	1	0	1
5	5	7	BOGO	9b98b8c7a33c4b65b9aebfe6a799e6d9	1	1	0	1
5	20	10	Discount	0b1e1539f2cc45b7b9fa7c272da2e1d7	1	0	0	1

Profile Dataset: ===================================================================================

Gender	Age	ID	Became Member On	Income	Member Since Year	Member Tenure
Other	55	68be06ca386d4c31939f3a4f0e3dd783	2017-02-12	64000	2017	23
Female	55	0610b486422d4921ae7d2bf64640c50b	2017-07-15	112000	2017	18
Other	55	38fe809add3b4fcf9315a9694bb96ff5	2018-07-12	64000	2018	6
Female	75	78afa995795e4d85b5d9ceeca43f5fef	2017-05-09	100000	2017	20
Other	55	a03223e636434f42ac4c3df47e8bac43	2017-08-04	64000	2017	17

Transcript Dataset: ================================================================================

Person	Event	Value	Time	Amount	Reward	Offer ID
78afa995795e4d85b5d9ceeca43f5fef	Offer Received	{'offer id': '9b98b8c7a33c4b65b9aebfe6a799e6d9'}	0	nan	nan	9b98b8c7a33c4b65b9aebfe6a799e6d9
a03223e636434f42ac4c3df47e8bac43	Offer Received	{'offer id': '0b1e1539f2cc45b7b9fa7c272da2e1d7'}	0	nan	nan	0b1e1539f2cc45b7b9fa7c272da2e1d7
e2127556f4f64592b11af22de27a7932	Offer Received	{'offer id': '2906b810c7d4411798c6938adc9daaa5'}	0	nan	nan	2906b810c7d4411798c6938adc9daaa5
8ec6ce2a7e7949b1bf142def7d0e0586	Offer Received	{'offer id': 'fafdcd668e3743c1bb461111dcafc2a4'}	0	nan	nan	fafdcd668e3743c1bb461111dcafc2a4
68617ca6246f4fbc85e91a2a49552598	Offer Received	{'offer id': '4d5c57ea9a6940dd891ad53e9dbe8da0'}	0	nan	nan	4d5c57ea9a6940dd891ad53e9dbe8da0

====================================================================================================

Feature Description in the Datasets

Portfolio Dataset

Feature	Description
Reward (int)	Given reward for completing an offer
Channels (list of strings)	Email, mobile app, social media, etc
Difficulty (int)	Minimum spending requirement for completing an offer
Duration (int)	Time that an offer is valid
Offer_Type (string)	Type of offer
ID (string)	Offer ID

Profile Dataset

Feature	Description
Gender (str)	Customers gender
Age (int)	Customers age
ID (str)	Customers ID
Became_Member_On (int)	Date of membership
Income (float)	Customer's income

Transcript Dataset

Feature	Description
Person (str)	Customer ID
Event (str)	Record description
time (int)	Time in hours (since the beginning of the study)
Value - (dict of strings)	Offer ID or transaction amount

Exploratory Data Analysis

In this section, a variety of plots are provided to a better understanding of the relationships between features.

Channel Distribution for Various Offers

Group = Portfolio.groupby('Offer Type')[['Email','Mobile','Social','Web']].sum()
Group = Group.reset_index().melt(id_vars= ['Offer Type'], value_vars= Group.columns, var_name='Channels', value_name='Count')
Group ['Percentage'] = np.round(100* Group['Count'].values /Group['Count'].sum(), 2)
fig = px.bar(Group, x='Offer Type', y='Count', orientation='v',
             color = 'Channels', barmode='group',
             color_discrete_sequence= ['LightBlue', 'DeepSkyBlue', 'CornFlowerBlue', 'RoyalBlue', 'MediumBlue', 'Navy'],
             text = 'Percentage', width = 800, height= 500)
fig.update_traces(marker_line_color= 'Black', marker_line_width=0.8, opacity=1,
                  texttemplate=' %{text:.2}%', textposition='inside')
fig.update_xaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=False, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=False, gridwidth=1, gridcolor='Lightgray')
fig.update_yaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=True, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=True, gridwidth=1, gridcolor='Lightgray', range=[0, 5])
fig.update_layout(plot_bgcolor= 'white', title={'text': '<b>' + 'Channel Distribution for Various Offers' + '<b>',
                         'x':0.5, 'y':0.91, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()
del Group

It seems that social media is the least efficient channel among all.

Age Distribution

def FeatDispPlot(Feat, nbins = 10, Color = 'GreenYellow', LC = 'Black',
                 xLim = None, yLim = [0, 6e3], H = 450, titleY = 0.92, Inp = Profile):
    fig = px.histogram(Inp, x = Feat, nbins=nbins, marginal= 'box',
                       color_discrete_sequence= [Color], hover_data= [Feat])
    fig.update_xaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                     zeroline=False, zerolinewidth=1, zerolinecolor='Black',
                     showgrid=False, gridwidth=1, gridcolor='Lightgray')
    fig.update_yaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                     zeroline=True, zerolinewidth=1, zerolinecolor='Black',
                     showgrid=True, gridwidth=1, gridcolor='Lightgray')
    # Median
    fig.add_trace(go.Scatter(x= Inp[Feat].median()* np.ones(int(yLim[1])), y= np.arange(int(yLim[0]), int(yLim[1])),
                         name="Median", line=dict(color='RoyalBlue', width=2, dash='dot')))
    # Mean
    fig.add_trace(go.Scatter(x= Inp[Feat].mean()* np.ones(int(yLim[1])), y= np.arange(int(yLim[0]), int(yLim[1])),
                         name="Mean", line=dict(color='Red', width=2, dash='dot')))

    Name = '%s Distribution' % Feat
    fig.update_layout(legend_orientation='v', plot_bgcolor= 'white', height= H, width= 900,
                     title={'text': '<b>' + Name + '<b>', 'x':0.5, 'y': titleY, 'xanchor': 'center', 'yanchor': 'top'},
                     yaxis_title='Frequency')
    fig.update_traces(marker_line_color= LC, marker_line_width=0.5, opacity=1)
    fig['layout']['yaxis'].update(range=yLim)
    if not xLim  == None:
        fig['layout']['xaxis'].update(range=xLim)
    fig.show()

FeatDispPlot('Age')

Costumers are mainly from the age group 50-60.

Income Distribution

FeatDispPlot('Income', Color = 'LightSkyblue', xLim = [2e4, 14e4], yLim = [0, 5e3])

Most of the customers' income is around 60K.

Gender Distribution

def FeatDispPlot2(Feat, LC = 'Black',
                  Colors =  ['HoneyDew','GreenYellow', 'Bisque','LightSalmon','Plum', 'LightSkyBlue'],
                  yLim = [0, 9e3], H = 450, titleY = 0.92, Inp = Profile):
    Group = Inp.groupby([Feat])[Feat].agg({'count'}).rename(columns = {'count':'Count'})
    Group['Percentage'] = np.round(100* Group.values /Group.sum().values, 2)
    Group.reset_index(drop = False, inplace = True)
    
    fig = make_subplots(rows=1, cols=2, specs=[[{"type": "xy"}, {'type':'domain'}]])
    # Left
    fig.add_trace(go.Bar(x= Group[Feat].values, y= Group['Count'].values, marker_color= Colors,
                         textposition='inside', showlegend = False, hovertext=list(Group['Percentage'])), 1, 1)
    fig.update_layout(plot_bgcolor= 'white')
    fig.update_traces(marker_line_color= LC, marker_line_width=1, opacity=1, row=1, col=1)
    fig.update_xaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=False, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=False, gridwidth=1, gridcolor='Lightgray')
    fig.update_yaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=True, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=True, gridwidth=1, gridcolor='Lightgray', range= yLim)
    # Right
    fig.add_trace(go.Pie(labels= Group[Feat].values, values= Group['Count'].values, textfont=dict(size=16),
                     marker=dict(colors = Colors, line=dict(color='black', width=1))), 1, 2)
    fig.update_traces(marker_line_color= LC, marker_line_width=1, opacity=1, row=1, col=2)
    fig.update_layout(plot_bgcolor= 'white', width = 900,
                  title={'text': '<b>' + '%s Distribution' % Feat + '<b>',
                         'x':0.5, 'y':0.90, 'xanchor': 'center', 'yanchor': 'top'})
    fig.show()

FeatDispPlot2('Gender', Colors =  ['LightPink','CornFlowerBlue','LightGreen'])

The number of male customers is larger than the number of female customers.

Memberships

FeatDispPlot2('Member Since Year', yLim = [0, 7e3])

Membership Growth Over Year by Gender

Group = Profile.groupby(['Member Since Year','Gender'])['Gender'].agg({'count'}).rename(columns ={'count':'Count'})
Group['Percentage'] = np.round(100* Group.values /Group.sum().values, 2)
Group.reset_index(drop = False, inplace = True)
Group['Member Since Year'] = Group['Member Since Year'].astype(str)
Group.columns = [x.replace('_',' ') for x in Group.columns]
Colors =  ['LightPink','CornFlowerBlue','LightGreen']

fig = px.bar(Group, y= 'Member Since Year', x= 'Percentage', orientation='h',
             color = 'Gender', text = 'Percentage', color_discrete_sequence= Colors, hover_data = Group.columns, height= 450)
fig.update_traces(marker_line_color=  'Navy', marker_line_width=1, opacity=1)
fig.update_traces(texttemplate='%{text:.2}%', textposition='inside')
fig.update_layout(plot_bgcolor= 'white')
fig.update_traces(marker_line_color= 'Black', marker_line_width=1, opacity=1, row=1, col=1)
fig.update_xaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=True, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=True, gridwidth=1, gridcolor='Lightgray', range=[0, 40])
fig.update_yaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=False, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=False, gridwidth=1, gridcolor='Lightgray')
fig.update_layout(uniformtext_minsize= 8, uniformtext_mode='hide')
fig.update_layout(plot_bgcolor= 'white', width = 950,
                  title={'text': '<b>' + 'Membership Growth Over Year by Gender' + '<b>',
                         'x':0.5, 'y':0.95, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()
del Group

Assuming the data is a snapshot of the end of 2018. get member tenure by the number of months.

Group = Profile[['Member Tenure','Gender']].dropna()
Group.columns = [x.replace('_',' ') for x in Group.columns]

Colors =  ['LightPink','CornFlowerBlue','LightGreen']
fig = px.histogram(Group, x = 'Member Tenure', color='Gender', nbins=100,  marginal= 'box',  color_discrete_sequence= Colors,
                  hover_data= Group.columns)
fig.update_layout(plot_bgcolor= 'white')
fig.update_traces(marker_line_color= 'Navy', marker_line_width=1, opacity=1)
fig.update_xaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True)
fig.update_yaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 showgrid=True, gridwidth=1, gridcolor='Lightgray')
fig['layout']['yaxis'].update(range=[0, 1000])
fig['layout']['xaxis'].update(range=[Group['Member Tenure'].min()-1, Group['Member Tenure'].max()+1])
fig.update_layout(plot_bgcolor= 'white', width = 950,
                  title={'text': '<b>' + 'Member Tenure Distribution by Gender' + '<b>',
                         'x':0.5, 'y':0.95, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()
del Group

Event Plots

Summary_Transcript = pd.DataFrame(Transcript.groupby(['Event'])['Event'].count())
Summary_Transcript.columns = ['Count']
display(Summary_Transcript.T)
data = {'Completed': [100*Summary_Transcript.loc['Offer Completed','Count']/Summary_Transcript.loc['Offer Viewed','Count']],
        'Viewed':[100*Summary_Transcript.loc['Offer Viewed','Count']/Summary_Transcript.loc['Offer Received','Count']]}
Group = pd.DataFrame(data = data).round(2).T
Group.columns = ['Percentage']
display(Group.T)

fig = plt.figure(figsize=(15, 6), constrained_layout=True)
gs = fig.add_gridspec(1, 5)
ax0 = fig.add_subplot(gs[:-1])
ax1 = fig.add_subplot(gs[-1])
_ = sns.barplot(ax = ax0, y="Event", x="Count", palette='Purples',
                edgecolor='k',  hatch="///", data=Summary_Transcript.reset_index(drop = False))
_ = ax0.set_xlim([0, 14e4])
_ = ax0.set_xlabel('Count', fontsize = 14)

_ = sns.barplot(ax = ax1, x="index", y="Percentage", palette='Blues',
                edgecolor='k',  hatch="///", data=Group.reset_index(drop = False))
_ = ax1.set_xlabel(None, fontsize = 14)
_ = ax1.set_ylim([0,100])
_ = ax1.set_yticks(np.arange(0, 101, 10))

plt.subplots_adjust(wspace= 0.5)

Event	Offer Completed	Offer Received	Offer Viewed	Transaction
Count	33579	76277	57725	138953

	Completed	Viewed
Percentage	58.17	75.68

Group = Data.groupby(['Event','Offer Type'])['Offer Type'].agg({'count'}).rename(columns ={'count':'Count'})
Group['Percentage'] = np.round(100* Group.values /Group.sum().values, 2)
Group.reset_index(drop = False, inplace = True)
Colors =   ['GreenYellow','LightSalmon','LightSkyBlue']

fig = px.bar(Group, y= 'Offer Type', x= 'Percentage', orientation='h',
             color = 'Event', text = 'Percentage', color_discrete_sequence= Colors, hover_data = Group.columns, height= 350)
fig.update_traces(marker_line_color=  'Navy', marker_line_width=1, opacity=1)
fig.update_traces(texttemplate='%{text:.2}%', textposition='inside')
fig.update_layout(uniformtext_minsize= 8, uniformtext_mode='hide')
fig.update_layout(plot_bgcolor= 'white')
fig.update_xaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=True, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=True, gridwidth=1, gridcolor='Lightgray', range=[0, 50])
fig.update_yaxes(showline=True, linewidth=1, linecolor='Lightgray', mirror=True,
                 zeroline=False, zerolinewidth=1, zerolinecolor='Black',
                 showgrid=False, gridwidth=1, gridcolor='Lightgray', title_text='Offer Type')
fig.update_layout(plot_bgcolor= 'white', width = 950,
                  title={'text': '<b>' + 'Event by Offer Type' + '<b>',
                         'x':0.5, 'y':0.92, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()
del Group

A Customer's Journey

First off,

fig = go.Figure(go.Sankey(arrangement = "snap",
                          node = {"label": ['Offer Received', 'Offer Viewed', 'Transaction', 'Offer Completed'],
                                  "x": [0, 0, 0, 0], "y": [0, 1, 2, 3], 'pad':10},
                          link = {"source": [0, 1, 2, 3], "target": [1, 2, 3], "value": [.5, .5, .5, .5]}))
fig.update_layout(height = 200, title={'text': '<b>' + """A Customer's Journey""" + '<b>',
                         'x':0.5, 'y':0.92, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()

Now, let's focus on those customers that completed an offer.

# creating a list of these customers
Ofer_Completed_List = Data.loc[Data['Event'] == 'Offer Completed', 'Person'].tolist()

The journey of one of these customers can be analyzed as well. For example, the first customer from the above list

Customer_Journey = Data.loc[Data.Person == Ofer_Completed_List[0],:].sort_index()
Customer_Journey.head(6).style.hide_index()

Person	Event	Time	Amount	Reward Received	Offer ID	Reward Defined	Difficulty	Duration	Offer Type	Email	Mobile	Social	Web
9fa9ae8f57894cc9a3b8a9bbe0fc1b2f	Offer Received	0	nan	nan	2906b810c7d4411798c6938adc9daaa5	2.000000	10.000000	7.000000	Discount	1.000000	1.000000	0.000000	1.000000
9fa9ae8f57894cc9a3b8a9bbe0fc1b2f	Offer Viewed	0	nan	nan	2906b810c7d4411798c6938adc9daaa5	2.000000	10.000000	7.000000	Discount	1.000000	1.000000	0.000000	1.000000
9fa9ae8f57894cc9a3b8a9bbe0fc1b2f	Transaction	0	34.560000	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan
9fa9ae8f57894cc9a3b8a9bbe0fc1b2f	Offer Completed	0	nan	2.000000	2906b810c7d4411798c6938adc9daaa5	2.000000	10.000000	7.000000	Discount	1.000000	1.000000	0.000000	1.000000
9fa9ae8f57894cc9a3b8a9bbe0fc1b2f	Transaction	42	21.550000	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan
9fa9ae8f57894cc9a3b8a9bbe0fc1b2f	Transaction	114	32.870000	nan	nan	nan	nan	nan	nan	nan	nan	nan	nan

Feature Difference by Offer Combination

To have plots for this part. We need to calulate transactions amount without any offers and The number of transaction without any offers.

# Transactions amount without any offers
User_Data['Tran Amnt No Offer'] = User_Data['Tot Tran Amnt']- User_Data['Offer Trans Amnt']
# The number of transaction without any offers
User_Data['Trans Cnt No Offer'] = User_Data['Tot Tran Cnt']- User_Data['Offer Completed']
# Converting to integers
User_Data[['BOGO offer','Disc offer','Info offer']] = User_Data[['BOGO offer','Disc offer','Info offer']].astype(int)

Metric Average of Offer Type Combinations plots

In the following tables and plots, (BOGO Offer, Discount Offer, Informational Offer) shows that what offer has been used. For example, (1,1,0) means people who respond to BOGO Offers and Discount Offers but not Informational Offers.

def Mean_by_Offer_Type (Columns):
    Cols = ['BOGO offer','Disc offer','Info offer']
    j = Columns[0]
    Group = User_Data.groupby(Cols)[j].mean().reset_index()
    for j in Columns[1:]:
        Temp = User_Data.groupby(Cols)[j].mean().reset_index()
        Group = Group.merge(Temp, on =Cols)
        del Temp
    return Group

Column_List = ['Offer Tran Cnt Ratio', 'Offer Trans Amnt Ratio', 'Offer Comp View Ratio', 'Offer Comp Rec Ratio']

Group = Mean_by_Offer_Type(Column_List).set_index(['BOGO offer','Disc offer','Info offer'])
display(Group.round(4))

Colors =  ['GreenYellow', 'Bisque','LightSalmon','Plum', 'LightSkyBlue']
fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(16, 7), sharex=False)
_ = Group.plot.bar(ax = ax, rot = 0, edgecolor = 'Navy', color = Colors, hatch = '//')
_ = ax.set_title('Metric Average of Offer Type Combinations')
_ = ax.legend([Feat_Dict[x] for x in Column_List], loc='upper left', fontsize = 12)
_ = ax.set_xlabel('(BOGO Offer, Discount Offer, Informational Offer)')
_ = ax.set_ylim([0,1])

			Offer Tran Cnt Ratio	Offer Trans Amnt Ratio	Offer Comp View Ratio	Offer Comp Rec Ratio
BOGO offer	Disc offer	Info offer
0	0	0	0.0000	0.0000	0.0000	0.0000
		1	0.2110	0.3135	0.2807	0.2892
	1	0	0.2383	0.4033	0.5249	0.3604
		1	0.3114	0.5239	0.5364	0.5888
1	0	0	0.2728	0.3540	0.5516	0.3782
		1	0.3542	0.5111	0.5501	0.6056
	1	0	0.3840	0.4542	0.7589	0.6470
		1	0.4140	0.5511	0.6838	0.7924

Column_List = ['Offer Completed', 'Tot Tran Cnt', 'Trans Cnt No Offer', 'Tot Rewards Rec', 'Reward per Offer',
               'Difficulty per Offer']
Feat_Dict.update({'Offer Completed': 'Offer Completed', 'Trans Cnt No Offer':'Trans Count without Offer'})
Group = Mean_by_Offer_Type (Column_List).set_index(['BOGO offer','Disc offer','Info offer'])
display(Group.round(4))

Colors =  ['Blue', 'Orange', 'DarkGreen','Purple','LightPink', 'Crimson']

fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(16, 7), sharex=False)
_ = Group.plot.bar(ax = ax, rot = 0, edgecolor = 'Navy', color = Colors)
_ = ax.set_title('Metric Average of Offer Type Combinations')
_ = ax.legend([Feat_Dict[x] for x in Column_List], loc='upper left', fontsize = 12)
_ = ax.set_xlabel('(BOGO Offer, Discount Offer, Informational Offer)')
_ = ax.set_ylim([0,20])
_ = ax.set_yticks(np.arange(0, 21, 2))

			Offer Completed	Tot Tran Cnt	Trans Cnt No Offer	Tot Rewards Rec	Reward per Offer	Difficulty per Offer
BOGO offer	Disc offer	Info offer
0	0	0	0.0000	4.4779	4.4779	1.5248	0.0000	0.0000
		1	1.1944	7.5111	6.3168	1.6380	1.4593	0.0000
	1	0	1.4905	9.0246	7.5341	7.1875	5.1920	9.5161
		1	2.6759	10.8667	8.1909	6.6585	2.5487	5.2188
1	0	0	1.5340	7.7338	6.1998	15.7631	10.7039	7.2817
		1	2.7579	9.5486	6.7907	14.9018	5.4684	3.9611
	1	0	2.9435	9.8640	6.9205	18.6081	6.5097	8.8200
		1	3.9506	11.2950	7.3444	16.6150	4.2396	6.2400

Column_List = [ 'Tot Tran Amnt', 'Offer Trans Amnt', 'Tran Amnt per Offer', 'Ave Tran Amnt', 'Tran Amnt No Offer']
Feat_Dict.update({'Tran Amnt No Offer': 'Transaction Amount without Offer'})
Group = Mean_by_Offer_Type (Column_List).set_index(['BOGO offer','Disc offer','Info offer'])
display(Group.round(4))

fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(16, 7), sharex=False)
_ = Group.plot.bar(ax = ax, rot = 0, edgecolor = 'k',
    color = sns.set_palette(["#9b59b6", "#3498db", "#e74c3c", "#34495e","#2ecc71","#95a5a6"]))
_ = ax.set_title('Metric Average of Offer Type Combinations')
_ = ax.legend([Feat_Dict[x] for x in Column_List], loc='upper left', fontsize = 12)
_ = ax.set_xlabel('(BOGO Offer, Discount Offer, Informational Offer)')
_ = ax.set_ylim([0,200])
_ = ax.set_yticks(np.arange(0, 201, 20))

			Tot Tran Amnt	Offer Trans Amnt	Tran Amnt per Offer	Ave Tran Amnt	Tran Amnt No Offer
BOGO offer	Disc offer	Info offer
0	0	0	26.2267	0.0000	0.0000	6.1674	26.2267
		1	46.7299	14.4213	11.4998	7.2478	32.3086
	1	0	95.1184	29.5337	19.3329	13.3275	65.5847
		1	115.7472	52.0649	18.9348	12.5913	63.6824
1	0	0	120.0015	32.7871	21.4729	17.6009	87.2144
		1	146.2316	66.4001	23.5046	16.7318	79.8316
	1	0	175.0637	66.7588	22.2873	19.7209	108.3049
		1	187.4890	93.1559	23.3329	17.9300	94.3331

• We can see from the above plot that the customers who have taken advantage of all three offer types not only spend the most with offers but also have the highest total transaction amount. In other words, the higher amount of incentivization is, the more is spent.

• The BOGO offer is the leading offer in terms of having a higher total transaction amount than discount and informational offers. Besides, customers tend to spend more than half of the total transaction amount without offers while using BOGO and discount offers.

We would like to know how the overall offer completion rate, transaction amount ratio motivated by an offer, and reward per offer affected by the user demographics.

Offer Completion, Offer Transaction Amount and Reward Per Offer by Gender

def PlotTable(Feat1, Feat2, Inp = User_Data):
    Female_Avg = User_Data[User_Data['Gender Female']==1].groupby(Feat1).mean()[Feat2]
    Male_Avg = User_Data[User_Data['Gender Male']==1].groupby(Feat1).mean()[Feat2]
    return Female_Avg, Male_Avg
fig, ax = plt.subplots(nrows=3, ncols=3, figsize=(14, 13.5))
ax = ax.ravel()

# First Row
R = ['Offer Comp Rec Ratio', 'Offer Trans Amnt Ratio', 'Reward per Offer'] 
for i in range(len(R)):
    M, F = PlotTable('Member Tenure', R[i])
    _ = ax[i].scatter(F.index, F, label='Female', color = 'tomato', edgecolor = 'darkred')
    _ = ax[i].scatter(M.index, M, label='Male', color = 'steelblue', edgecolor = 'navy')
    _ = ax[i].set_xlim(0,70)
    _ = ax[i].set_xlabel('Member Tenure', fontsize = 11)
    _ = ax[i].set_ylabel('Avg ' + Feat_Dict[R[i]], fontsize = 11)
    _ = ax[i].set_title('Avg ' + Feat_Dict[R[i]] + '\n' + 'By Membership Tenure', fontsize = 11)
    _ = ax[i].legend(loc='lower right')
    _ = ax[i].set_xlim(0, 70)
    _ = ax[i].set_ylim(0, np.ceil(max(M.max(),F.max())))
    
    
# Second Row
for i in range(len(R)):
    M, F = PlotTable('Income', R[i])
    _ = ax[3+i].scatter(F.index, F, label='Female', color = 'tomato', edgecolor = 'darkred')
    _ = ax[3+i].scatter(M.index, M, label='Male', color = 'steelblue', edgecolor = 'navy')
    _ = ax[3+i].set_xlim(0,70)
    _ = ax[3+i].set_xlabel('Income', fontsize = 11)
    _ = ax[3+i].set_ylabel('Avg ' + Feat_Dict[R[i]], fontsize = 11)
    _ = ax[3+i].set_title('Avg ' + Feat_Dict[R[i]] + '\n' + 'By Membership Tenure', fontsize = 11)
    _ = ax[3+i].legend(loc='lower right')
    _ = ax[3+i].set_xlim(0, 1.5e5)
    _ = ax[3+i].set_ylim(0, np.ceil(max(M.max(),F.max())))
    
# Third Row
for i in range(len(R)):
    M, F = PlotTable('Age', R[i])
    _ = ax[6+i].scatter(F.index, F, label='Female', color = 'tomato', edgecolor = 'darkred')
    _ = ax[6+i].scatter(M.index, M, label='Male', color = 'steelblue', edgecolor = 'navy')
    _ = ax[6+i].set_xlim(0,70)
    _ = ax[6+i].set_xlabel('Age', fontsize = 11)
    _ = ax[6+i].set_ylabel('Avg ' + Feat_Dict[R[i]], fontsize = 11)
    _ = ax[6+i].set_title('Avg ' + Feat_Dict[R[i]] + '\n' + 'By Membership Tenure', fontsize = 11)
    _ = ax[6+i].legend(loc='lower right')
    _ = ax[6+i].set_xlim(0, 120)
    _ = ax[6+i].set_ylim(0, np.ceil(max(M.max(),F.max())))
    
_ = plt.tight_layout()

Offer Conversion Rate

User_Data['BOGO Conv Rate'] = User_Data['BOGO comp']/User_Data['BOGO Offer Rec']
User_Data['Disc Conv Rate'] = User_Data['Disc comp']/User_Data['Disc Offer Rec']
User_Data['Info Conv Rate'] = User_Data['Info comp']/User_Data['Info Offer Rec']

Group = User_Data[['BOGO Conv Rate', 'Disc Conv Rate','Info Conv Rate']].agg({'mean'}).T

TempKey = ['BOGO Conv Rate', 'Disc Conv Rate', 'Info Conv Rate']
TempVals = ['BOGO Offer Conversion Rate', 'Discount Offer Conversion Rate', 'Information Offer Conversion Rate']
Temp = dict(zip(TempKey,TempVals))
Feat_Dict.update(Temp)
del Temp, TempKey, TempVals

fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(14, 6), sharex=False)
_ = Group.plot.bar(ax = ax[0], legend = False, color= 'mediumorchid', edgecolor='indigo', hatch = '///', fontsize = 14,rot=0)
_ = ax[0].set_xticks(np.arange(3))
_ = ax[0].set_xticklabels(['BOGO Offer\nConversion Rate', 'Information Offer\nConversion Rate',
                            'Discount Offer\nConversion Rate'])
_ = ax[0].set_title('Ave Conversion Rate by Offer', fontsize = 14)
_ = ax[0].set_xlabel('Average', fontsize = 14)
_ = ax[0].set_ylim([0, .5])
_ = ax[0].set_yticks(np.arange(0, .55, 0.05))

_ = Group.plot.pie(ax = ax[1], y= 'mean', startangle=90, label = '', labels = None,
                  colors = sns.set_palette(['lightskyblue','limegreen','lightsalmon']),
                  legend=True, autopct='%1.1f%%', fontsize=14,
                  pctdistance=0.85, explode = (0.05,0.05,0.05))
_ = ax[1].legend(bbox_to_anchor=(.25, 1), labels= [Feat_Dict[x] for x in Group.index], fontsize = 12)
_ = ax[1].add_artist(plt.Circle((0,0),0.70,fc='white'))

def PlotTable(Feat1, Feat2, Inp = User_Data):
    Female_Avg = User_Data[User_Data['Gender Female']==1].groupby(Feat1).mean()[Feat2]
    Male_Avg = User_Data[User_Data['Gender Male']==1].groupby(Feat1).mean()[Feat2]
    return Female_Avg, Male_Avg
fig, ax = plt.subplots(nrows=3, ncols=3, figsize=(14, 13.5))
ax = ax.ravel()

# First Row
R = ['BOGO Conv Rate', 'Disc Conv Rate', 'Info Conv Rate'] 
for i in range(len(R)):
    M, F = PlotTable('Member Tenure', R[i])
    _ = ax[i].scatter(F.index, F, label='Female', color = 'tomato', edgecolor = 'darkred')
    _ = ax[i].scatter(M.index, M, label='Male', color = 'steelblue', edgecolor = 'navy')
    _ = ax[i].set_xlim(0,70)
    _ = ax[i].set_xlabel('Member Tenure', fontsize = 11)
    _ = ax[i].set_ylabel('Avg ' + Feat_Dict[R[i]], fontsize = 11)
    _ = ax[i].set_title('Avg ' + Feat_Dict[R[i]] + '\n' + 'By Membership Tenure', fontsize = 11)
    _ = ax[i].legend(loc='lower right')
    _ = ax[i].set_xlim(0, 70)
    _ = ax[i].set_ylim(0, np.ceil(max(M.max(),F.max())))
    
    
# Second Row
for i in range(len(R)):
    M, F = PlotTable('Income', R[i])
    _ = ax[3+i].scatter(F.index, F, label='Female', color = 'tomato', edgecolor = 'darkred')
    _ = ax[3+i].scatter(M.index, M, label='Male', color = 'steelblue', edgecolor = 'navy')
    _ = ax[3+i].set_xlim(0,70)
    _ = ax[3+i].set_xlabel('Income', fontsize = 11)
    _ = ax[3+i].set_ylabel('Avg ' + Feat_Dict[R[i]], fontsize = 11)
    _ = ax[3+i].set_title('Avg ' + Feat_Dict[R[i]] + '\n' + 'By Membership Tenure', fontsize = 11)
    _ = ax[3+i].legend(loc='lower right')
    _ = ax[3+i].set_xlim(0, 1.5e5)
    _ = ax[3+i].set_ylim(0, np.ceil(max(M.max(),F.max())))
    
# Third Row
for i in range(len(R)):
    M, F = PlotTable('Age', R[i])
    _ = ax[6+i].scatter(F.index, F, label='Female', color = 'tomato', edgecolor = 'darkred')
    _ = ax[6+i].scatter(M.index, M, label='Male', color = 'steelblue', edgecolor = 'navy')
    _ = ax[6+i].set_xlim(0,70)
    _ = ax[6+i].set_xlabel('Age', fontsize = 11)
    _ = ax[6+i].set_ylabel('Avg ' + Feat_Dict[R[i]], fontsize = 11)
    _ = ax[6+i].set_title('Avg ' + Feat_Dict[R[i]] + '\n' + 'By Membership Tenure', fontsize = 11)
    _ = ax[6+i].legend(loc='lower right')
    _ = ax[6+i].set_xlim(0, 120)
    _ = ax[6+i].set_ylim(0, np.ceil(max(M.max(),F.max())))
    
_ = plt.tight_layout()

---